Forced and voluntary aerobic cycling interventions improve walking capacity in individuals with chronic stroke

Abstract

Objectives:

To determine the efficacy of high-intensity cycling to improve walking capacity in persons with chronic stroke, identify variables that predict improvement in walking capacity, and quantify the relationship between Six Minute Walk Test (6MWT) and Cardiopulmonary Exercise (CPX) Test variables.

Design:

Secondary analysis of data from 2 randomized clinical trials

Setting:

Research laboratory

Participants:

Individuals with chronic stroke (N=43)

Interventions:

Participants were randomized to one of the following time-matched interventions, occurring 3 times per week for 8 weeks: 1) forced aerobic exercise and upper extremity repetitive task practice (FE+RTP, N=16), 2) voluntary aerobic exercise and upper extremity repetitive task practice (VE+RTP, N=14), or 3) a non-aerobic control group (control, N=13).

Main Outcome Measure:

Change in walking capacity as measured by the 6MWT from baseline to end of treatment (EOT).

Results:

Significant increases were observed in distance traveled during the 6MWT at EOT compared to baseline in the FE+RTP (p<0.001) and VE+RTP (p<0.001) groups, but not in the control group (p=0.21). Among aerobic exercise participants, a multivariate regression analysis revealed that cycling cadence, power output, and baseline 6MWT distance were significant predictors of change in walking capacity.

Conclusions:

An 8-week aerobic cycling intervention prescribed at 60–80% of heart rate reserve and moderate to high cadence and resistance led to significant improvements in walking capacity in our cohort of persons with chronic stroke. Individuals with low baseline levels of walking capacity may benefit most from aerobic cycling to improve over ground locomotion. While the 6MWT did not elicit a comparable cardiorespiratory response as the maximal exertion CPX test, the 6MWT can be considered a valid and clinically relevant submaximal test of cardiorespiratory function in individuals with chronic stroke.

Stroke is a leading cause of severe, long-term disability among older adults in the United States.¹ The prevalence of stroke survivors with residual neurological deficits is approximately 6.4 million.^1,2 Despite advances in rehabilitation, nearly 75% of individuals do not regain full use of their hemiparetic lower extremities, resulting in deficits in locomotion, increased fall risk, and decreased community reintegration.^2–4 Considerable effort is put toward the recovery of gait post-stroke, with particular emphasis on walking speed, as higher walking speeds are predictive of reduced disability.⁵ Body-weight supported treadmill training and robotic-assisted training have been studied extensively over the past two decades. However, neither has been shown efficacious in improving walking speed or walking capacity, suggesting that methods of optimizing gait recovery are not well understood.⁶

The efficacy of aerobic exercise to improve gait outcomes has been explored, particularly with task-specific modalities such as overground or treadmill walking. Although walking and mixed aerobic interventions have been shown to improve Six-Minute Walk Test (6MWT) performance in chronic stroke, cycling interventions have shown less promising results.^5,7–9 A recently published clinical practice guideline found weak evidence for cycling interventions to improve locomotor function in persons with chronic stroke. However, of the studies referenced, those implementing high intensity protocols were found most efficacious, and further research was recommended to evaluate the effects of cycling on locomotor function.⁹

While the 6MWT has historically been used as a measure of cardiopulmonary fitness and as a common data element in clinical trials investigating the efficacy of aerobic exercise interventions, there is debate regarding what physiologic variables are captured by the 6MWT in the chronic stroke population.^10–12 In healthy adults and people with cardiorespiratory disease, the 6MWT has been shown to correlate with aerobic capacity;^13,14 however, impairments such as balance, spasticity, weakness, and diminished motor control have been shown to influence 6MWT performance in individuals post-stroke.^11,12,15,16 Additionally, survivors of stroke exhibit higher levels of aerobic capacity (VO_2peak) during cardiopulmonary exercise (CPX) testing than during the 6MWT, suggesting that the 6MWT does not tax the cardiorespiratory system comparably to CPX testing.^12,17,18 Given that physical therapists commonly utilize the 6MWT in lieu of a CPX for exercise prescription following stroke,^19–21 it is essential to understand the relationship between these two assessments.

The primary aim of this study was to determine the efficacy of high-intensity cycling to improve locomotor function. Secondary aims included identifying factors that predict improvement in 6MWT distance and quantifying the relationship between 6MWT and CPX test variables. Based on previous evidence that high-intensity aerobic exercise improves locomotor function,⁷ we hypothesized that those undergoing high-intensity cycling would demonstrate improved walking capacity compared to a non-aerobic control group. Additionally, we hypothesized that 6MWT would positively correlate with CPX variables.

MATERIALS AND METHODS

Cardiovascular data from two similar randomized controlled trials were analyzed for this study (clinicaltrials.gov registration numbers NCT02076776 funded by NIH and NCT02494518 funded by the American Heart Association). Both studies were approved by the Cleveland Clinic Institutional Review Board and all participants completed the informed consent process.

Participants

Individuals ≥ 6 months following a single, unilateral stroke with residual upper extremity (UE) hemiparesis were recruited for participation. Primary inclusion criteria were: 1)19–55 on the UE Motor section of the Fugl-Meyer Assessment, and 2)between 18–85 years of age. Primary exclusion criteria were: 1)hospitalization for myocardial infarction, congestive heart failure, or heart surgery within 3 months of study enrollment, 2)serious cardiac arrhythmia, 3)hypertrophic cardiomyopathy, 4)severe aortic stenosis, 5)uncontrolled hypertension, 6)pulmonary embolus, 7)other medical or musculoskeletal contraindication to exercise, 8)significant cognitive impairment, or 9)anti-spasticity injection in the upper extremity within 3 months of study enrollment. A CONSORT diagram has been published previously.²²

Outcomes of Interest

All participants underwent a CPX test on an upright cycle ergometer (Lode Excalibur Sport with Pedal Force Measurement, Lode B.V., Groningen, Netherlands) prior to randomization and at end of treatment (EOT). A 12-lead electrocardiogram was used to monitor heart rate and rhythm continuously five minutes prior to, during, and for five minutes following the CPX test. An incremental workload protocol was used beginning at 20 Watts (W) and increasing in 20 W stages every two minutes until attaining 100 W, at which time resistance was increased by 40 W every two minutes. The CPX test was terminated based on the American College of Sports Medicine maximal exertion criteria or volitionally by the participant due to fatigue.²³ A cardiologist blinded to group allocation interpreted the CPX test results to determine cardiopulmonary response to maximal exercise. Gas analyses, including volume of oxygen consumption, were averaged over the final 30 seconds of each stage to capture steady state. Peak volume of oxygen (VO_2peak) was defined as the highest averaged sample obtained.

The 6MWT was administered at baseline and EOT by a therapist blinded to group allocation. Participants were instructed to cover as much distance as safely possible in six minutes on a continuous rectangular course. All participants wore a gait belt for safety and used assistive devices (cane, quad cane, wheeled walker or rollator) and orthoses (ankle-foot orthosis or ankle brace) as they were accustomed to for walking outside of their home environment. Each participant used the same equipment (assistive device and bracing) at baseline and EOT. Heart rate (HR) was obtained prior to (while sitting) and immediately following the test.

Interventions

Following baseline testing, participants were randomized using an envelope pull method to the following: forced exercise and repetitive task practice (FE+RTP), voluntary exercise and RTP (VE+RTP), and a non-exercise RTP control group (control). All sessions were ~90 minutes in length, with participants attending three days/week for eight weeks. The FE+RTP and VE+RTP groups underwent 45 minutes of aerobic exercise (AE) using different cycling modes described below followed by 45 minutes of RTP. In lieu of AE, the control group completed either 90-minutes of UE RTP, or 45-minutes of stroke-related education followed by 45-minutes of UE RTP. For purposes of examining cardiovascular response to AE, both non-aerobic exercise groups were combined into one control group.

Forced and voluntary exercise Interventions:

Participants in the FE+RTP and VE+RTP groups exercised on semi-recumbent stationary bicycles at 60–80% of their heart rate reserve (HRR) for 45 minutes. Each 45-minute session included a 5-minute warm-up, 35-minute main exercise set, and 5-minute cool down. While the VE+RTP group exercised at a self-selected rate, the FE+RTP group exercised on a custom-engineered cycle equipped with a motor that augmented pedaling rate by 30% greater than the participant’s voluntary rate achieved during their baseline CPX test.²⁵ The FE mode was not passive as participants had to contribute in order to elevate their HR to achieve 60–80% of HRR. To facilitate adherence to the AE prescription, HR was continuously monitored using a chest strap and displayed for participants in both groups, who were verbally and visually cued to exercise within their target HR ranges. Following the FE or VE intervention, all participants completed 45 minutes of UE RTP.

Upper Extremity Repetitive Task Practice:

Repetitive task practice is the current clinical standard in stroke rehabilitation to facilitate the recovery of upper limb function.^26,27 The RTP approach was modeled after Lang and colleagues, and utilized motor learning principles emphasizing highly-repetitious blocked practice tasks that were functional, goal-oriented and relevant to the participant.²⁶ The RTP activities were administered by a neurologic physical therapist who tailored tasks to each individual ensuring appropriate difficulty and relevance.^22,28

Statistical Analysis

Descriptive statistics were computed to describe demographic variables and group characteristics. Randomized groups were compared on participant demographics, using ANOVA for normally distributed variables, the Kruskal-Wallis test for non-normally distributed continuous variables, or the chi-squared test for categorical variables. A linear mixed effects model was utilized at the 0.05 significance to determine the effect of group, time point, and their interaction on 6MWT as measured by the total distance traveled in meters. The model included random intercept and subject ID as a random effect. To check for normality and homoscedasticity assumptions for using a linear mixed effect model, the residual plots from the model were visually inspected to ensure no obvious deviation. A Wald Chi-squared test was used to determine if the factors, group and visit and their interaction were significant. Post-hoc pairwise comparisons were performed using least squares means (https://CRAN.R-project.org/package=LSMEANS) with a Bonferroni correction to control for type I errors.

Next, a step-wise multivariate linear regression was performed among exercisers to determine which demographic characteristics and exercise variables were associated with improved 6MWT performance. Participant characteristics included in the model were: age, group allocation, body mass index, sex, and baseline 6MWT distance. Exercise-related variables included in the model were aerobic exercise intensity (% HRR), exercise rate (pedaling cadence in RPM’s), power, and change in VO_2peak. To check for collinearity, variance inflation factors were examined. Finally, correlation coefficients were computed to explore the relationship between 6MWT and CPX outcomes. Significance was set at p<0.05 and statistical analyses were performed using R software version 3.4.0 (https://cran.r-project.org).

RESULTS

Participant demographics, baseline characteristics, exercise variables, and the primary outcomes are summarized in Table 1. Aerobic intensity and power were similar across groups; however, cadence was significantly higher for the FE+RTP group compared to VE+RTP (p=0.02).

Table 1.

Summary statistics presented as mean ± standard deviation for normally distributed data, median [Q1, Q3] for skew data, or N (%) for categorical data.

Demographic Variables	FE+RTP (N=16)		VE+RTP (N=14)		Control (N=13)		p-value
Age	51±12		60±14		59±11		0.07
Male sex (versus female)	12 (75%)		9 (64%)		12 (92%)		0.26
Body Mass Index	31.6 ± 6.6		28.6 ± 6.0		32.4 ± 8.5		0.34
Race:							0.16
African American	8 (50%)		2 (14%)		5 (38%)
White	5 (31%)		12 (86%)		7 (54%)
Asian	1 (6%)		0 (0%)		0 (0%)
Other	2 (12%)		0 (0%)		1 (8%)
Hispanic ethnicity	1 (6%)		0 (0%)		1 (8%)		0.16
Dominant side affected (%)	10 (62%)		7 (50%)		6 (46%)		0.59
Months since stroke	12 [7,16]		14 [11,28]		12 [9,20]		0.58
Baseline UE Fugl Meyer score	37±8		33±12		30±9		0.15
Baseline Peak VO2 (mL/kg/min)	18.0±6.1		17.0±3.8		15.7±3.2		0.42
Baseline Six Minute Walk Test (meters)	417 [347, 457]		295 [211, 384]		335 [186, 379]		0.02
Respiratory Exchange Ratio (Baseline)	1.15±0.07		1.18±0.08		1.13±0.10		0.77
Respiratory Exchange Ratio (EOT)	1.15±0.08		1.17±0.09		1.18±0.08		0.36
Exercise Variables
Average cadence (RPM)	74 [68, 82]		60 [56, 69]		--		0.02
Percentage of HRR (%)	60 [48, 65]		57 [46, 61]		--		0.52
Power (watts)	32.4 [24.1, 41.8]		35.8[31.7, 63.6]		--		0.16
Aerobic exercise time (min/session)	41.9±2.7		41.5±5.9		--		0.81
Outcome	FE + RTP		VE + RTP		Control		p-value for Interaction
	Baseline	EOT	Baseline	EOT	Baseline	EOT
Six Minute Walk Test (meters)	417 [348, 457]	480 [370, 515]	296 [211, 384]	327 [234, 432]	335 [186, 379]	312 [129,436]	0.07

Abbreviations: FE+RTP, forced-rate exercise and repetitive task practice; VE+RTP, voluntary-rate exercise and repetitive task practice; UE, upper extremity; VO₂, volume of oxygen consumption; mL/kg/min, milliliters per kilogram per minute; m, meters; EOT, End of Treatment; %HRR, percentage of heart rate reserve

Improved Walking Capacity Following Exercise

As depicted in Figure 1, the FE+RTP and VE+RTP groups improved in distance traveled during the 6MWT by a mean of 63 meters (15%) and 31 meters (10%), respectively, while the control group declined by a mean of 23 meters (7%). The linear mixed effect model revealed a significant effect of group on 6MWT distances (Chisq=7.7, p=0.02), a non-significant trend in the interaction effect of group by time point (Chisq=5.2, p=0.07), and a non-significant effect of time point (Chisq=1.8, p=0.18). Post hoc analysis determined a significant increase in distance traveled during the 6MWT at EOT compared to baseline in the FE+RTP (p<0.001) and VE+RTP (p<0.001) groups, but not in the control group (p=0.21). In addition, a within time point analysis determined a significant difference in measures between the FE+RTP and control group at baseline (p=0.04) and EOT (p<0.01), but no other significant differences were found.

Figure 1:

Change in Six Minute Walk Test from baseline to end of treatment for each FE (left panel), VE (middle panel) and control group (right panel) participant is depicted in light gray with black bold lines representing group means. To provide context, the 6MWT normative distance range for individuals 50–70 years of age is presented in light blue while categorizations of home versus community ambulatory capacity as defined by Fulk and colleagues are presented along the right y-axis. The greatest improvement in 6MWT distance was made by FE group participants with the group mean at EOT approaching age norm values, though a significant improvement was also evident for the VE group. Among the 30 exercisers, 6 (20%) improved to a higher ambulatory capacity category. Cycling Improves Walking Capacity Poststroke

Exercise cadence, Power, and Baseline Walking Capacity Predict Improvement in 6MWT Performance

As shown in Figure 2, considerable variability in change in 6MWT performance was observed among participants. The step-wise regression analysis with change in 6MWT performance as the response (F_3,26 = 2.88, p = 0.055, Multiple R squared = 0.25) found that baseline 6MWT distance, exercise cadence and power optimized the model, with lower baseline 6MWT distance, higher cadence, and greater power predictive of improved 6MWT performance. Variance inflation factors ranged from 1.2 to 2.9.

Figure 2:

Percent change in Six Minute Walk Test distance is depicted for each participant in the exercise (blue) and control (orange) groups. The solid vertical line demarcates those who improved by at least 10% versus those who did not. Higher exercise cadence and power were predictive of greatest improvements in walking capacity.

6MWT Performance Correlates with CPX Test Variables

As shown in Figures 3a and 3b, moderate correlations were observed between 6MWT distance and VO_2peak (R=0.68), and 6MWT distance and maximal watts achieved during the CPX test (R=0.65). Maximum HR achieved during the CPX test was compared to HR achieved at the end of the 6MWT, revealing a moderate correlation (Figure 3c, R=0.56). On average, HR achieved at the end of the 6MWT was 65% of maximum HR achieved during the CPX tests. Interestingly, change in 6MWT distance from baseline to EOT did not correlate with change in VO_2peak from baseline to EOT (Figure 3d, R=0.06).

Figure 3:

The relationship between Six Minute Walk Test (6MWT) and cardiopulmonary exercise (CPX) testing variables was compared revealing moderate correlations between 6MWT distance and peak oxygen consumption (Fig 3a), 6MWT distance and maximal power achieved during the CPX test (Fig 3b), and maximum heart rate response during the 6MWT and the CPX test (Fig 3c). A poor correlation was observed between change in 6MWT distance and change in peak oxygen consumption from baseline to end of treatment (Fig 3d).

DISCUSSION

Eight weeks of supervised forced- and voluntary-exercise cycling interventions improved walking capacity by 63 meters (15%) and 31 meters (10%), respectively, as measured by the 6MWT. Improvements by the FE group exceeded the minimal clinically important difference (MCID) for individuals with chronic stroke of 34.4 meters, while improvements by the VE group approached the MCID value.¹¹ These improvements in ambulatory capacity may lower the energy costs and functional burden associated with walking that are commonly experienced by individuals with chronic stroke.^10,16,29 Increased walking capacity has also been shown to reduce fall risk, improve community reintegration, and reduce long-term disability associated with stroke.^6,10 These meaningful improvements are unique from summary results provided in a recently published clinical practice guideline, in which weak evidence was reported for the efficacy of cycling interventions to improve locomotor function for individuals with chronic stroke.⁶ A number of reasons may explain the disconnect between the current results and recent practice guidelines. A limited number of studies (ie: five) used cycling as an intervention to inform the guidelines. Additionally, the majority of these protocols were classified as low aerobic intensity, and it was noted that cycling interventions performed at higher aerobic intensities, similar to the current protocol, may improve locomotor outcomes. Our results are encouraging demonstrating that with a specific exercise prescription and oversight, individuals with chronic stroke can achieve higher intensities of cycling both with and without assistance from the FE bike, resulting in significant improvements in locomotor function. The approach implemented in our protocol was in alignment with the strategies recommended by the guideline for clinical implementation, which included using devices that provide HR monitoring for feedback in real time, and providing target HR zones for patients.⁶

While cycling and walking have different motor control demands, both activities require the rapid reciprocal activation and relaxation of lower extremity muscles in a synergistic manner.^30–35 In a recent study, high speed cycling improved rate-dependent mobility as measured by the timed up and go in a cohort of healthy older adults.³⁶ In chronic hemiparesis, diminished power and abnormal timing and coordination of muscle agonists and antagonists disrupt the modulation of phasic muscle activity, thus resulting in inefficient movements.^3,30,32 While task-specific training has been considered critical in the recovery of gait post-stroke, the cycling intervention may have trained muscle groups to work synergistically to ensure smooth intra- and inter-limb reciprocal activation, similar to activation patterns used to coordinate joint angle accelerations and decelerations during various phases of the gait cycle.^35,37,38 Therefore, it is plausible that the training effect from our cycling protocol that entailed moderate to high exercise rate and aerobic intensity, resulted in improved walking capacity.

High Cadence and Power Result in Greatest Improvements in Walking Capacity

Considerable variability was observed in 6MWT performance from baseline to EOT among exercise groups. Therefore, we sought to determine which exercise and demographic variables were most predictive of improvements in 6MWT performance. Our model indicated that exercise cadence, power, and baseline 6MWT distance contributed most to predicting change in 6MWT performance, with the greatest improvements for those who trained at a higher cadence, higher power, and presented with lower baseline 6MWT values. Previous studies reported gait training at higher aerobic intensity and speed improves walking capacity, potentially by targeting central and peripheral deficits associated with stroke.^6,7 In addition to physiologic changes to muscle function attributed to the stroke itself, changes to peripheral tissue contribute to disability.³⁹ Cellular changes in skeletal muscle following stroke include altered fiber-type proportions, loss of type I muscle fibers, atrophy, and reductions in oxidative capacity³⁹. Physical activity and intensive exercise training have been shown to mitigate or reverse these physiologic changes, which in turn improve motor function, cardiovascular fitness, and metabolic health.^20,21,39 Thus, when considering parameters for exercise prescription post-stroke, challenging individuals to train at a relatively high cadence and resistance may be optimal to improve locomotor outcomes.

Not surprisingly, our cohort of individuals with chronic stroke presented with relatively low levels of ambulatory capacity and cardiovascular fitness at baseline, with a mean 6MWT distance of 319 meters and a mean VO_2peak of 16.8 mL/kg/min across all participants. Based on 6MWT ranges proposed by Fulk and colleagues to categorize locomotor status, 10 participants were classified as home ambulators, 6 were limited-community ambulators, and 27 were full-community ambulators at baseline.⁵ At post-intervention, 7 were categorized as home, 6 as limited-community, and 30 as full-community ambulators, potentially demonstrating a meaningful reduction in disability levels for at least 6 participants. These improvements in locomotor function as a result of aerobic cycling are encouraging, as it provides physical therapists with a rehabilitation approach that potentially transitions individuals from marginal levels of ambulatory capacity to achieving community levels of locomotion in a relatively short, 8-week timeframe.

Considerable debate exists regarding what is precisely measured by the 6MWT in individuals with chronic stroke.^12,14,16 Historically, the 6MWT has been proposed as a clinically feasible test of submaximal aerobic capacity.¹³ However, numerous studies have shown, and our own results have verified, that the motor limitations associated with post-stroke hemiplegia may prevent individuals from achieving activity levels during walking that adequately stress the cardiorespiratory system.^11,12,16 On average, the maximum HR attained during the 6MWT was only 65% of that attained during the CPX test, potentially demonstrating that our cohort of participants experienced greater motor limitations during locomotion compared to seated cycle ergometry utilized during the CPX test. Despite this disparity between activity modalities, the correlation between CPX and 6MWT variables implies that in our cohort of participants, both outcomes measure aspects of cardiopulmonary fitness, with the CPX obtaining maximal values while the 6MWT obtained submaximal values. Submaximal testing via the 6MWT has been suggested as a viable clinical alternative to maximal CPX testing in clinical settings where stress tests are not feasible, and submaximal exercise interventions are targeted.^20,21 An interesting finding was that changes in VO_2peak did not correlate with changes in 6MWT distance, suggesting that improvement in walking capacity is not solely influenced by improved aerobic fitness, but also by motor training associated with repetitive cycling movements.

Study Limitations

Levels of physical activity occurring outside of the study intervention were not monitored, yet may provide insight into change in locomotor function. Additionally, baseline 6MWT values were dissimilar across groups. Lastly, we did not obtain an objective biomechanical measure of lower extremity function, which may have influenced the multivariate regression model.

Conclusions

Our findings indicate that an 8-week aerobic cycling intervention prescribed at 60–80% of HRR and moderate to high cadence and resistance improves walking capacity in persons with chronic stroke. While cycling is not a task-specific intervention designed to improve locomotion, the intensive training may reverse peripheral changes in lower extremity skeletal muscle observed post-stroke or improve central mechanisms of motor control that translate into improvements in locomotion. Studies are ongoing using biomechanical gait assessments at baseline and post-intervention to more precisely delineate the effects of cycling on human gait in stroke survivors.

Abbreviations:

VO_2peak

peak oxygen uptake

AE

aerobic exercise

HRR

heart rate reserve

CPX

cardiopulmonary exercise

UE

upper extremity

FE+RTP

forced exercise and upper extremity repetitive task practice

VE+RTP

voluntary exercise and upper extremity repetitive task practice

6MWT

six-minute walk test

RPM

revolutions per minute

EOT

end of treatment

MCID

minimal clinically important difference